A scoping review on generative AI and large language models in mitigating medication related harm

Area of Science:

• Health Informatics
• Artificial Intelligence in Medicine
• Pharmacovigilance

Background:

• Medication-related harm significantly impacts global healthcare costs and patient outcomes.
• Generative artificial intelligence (GenAI) and large language models (LLMs) present potential solutions for mitigating these risks.

Purpose of the Study:

• To evaluate the scope and effectiveness of GenAI and LLMs in reducing medication-related harm.
• To identify key applications and assess the performance of these AI models in medication safety.

Main Methods:

• Systematic literature review across four databases.
• Inclusion criteria focused on studies published between January 2012 and October 2024.
• Analysis of 30 selected articles from an initial pool of 3988.

Main Results:

• GenAI and LLMs were applied in drug-drug interaction identification, clinical decision support, and pharmacovigilance.
• Models demonstrated potential in early adverse drug event identification, classification, and medication management support.
• Performance and utility varied across different applications.

Conclusions:

• GenAI and LLMs show promise for enhancing medication safety and reducing harm.
• Prospective studies are necessary to validate real-world integration and application of these AI technologies.
• Further investigation is required to fully understand the clinical utility and impact of AI in pharmacovigilance.